In the proposed study we explicitly recognize the importance of two types of social context (geographic and school context) for the health behavior decisions made by adolescents. Our guiding premise is that the characteristics of where adolescents live and go to school influence not only their behavioral alternatives, but also their associated social, economic and psychic costs. That is, we view these characteristics as both defining an opportunity structure that channels and constrains adolescent health behaviors and engendering social norms that delimit the boundaries of desirable or acceptable behaviors. In the proposed study, then, we will examine these contextual effects on the following five types of health behaviors: 1) sexual, contraceptive, and fertility behaviors; 2) alcohol, tobacco, and illicit drug use; 3) behaviors associated with physical injury; 4) exercise; and, 5) diet and eating disorders. Our first task during the proposed study is to construct a contextual database. This database will contain a complete set of aggregate measures for the places in which the adolescents included in the study spend the vast majority of their time: their schools, and their neighborhoods and communities. These aggregate measures will represent virtually all of the characteristics of these contexts that may be hypothesized to affect the health behaviors of adolescents. It will be a database that is necessary not only for our proposed analyses, but also for all of the other components of the larger study since those analyses explicitly recognize that school and geographic context constrain many of the individual- and group-processes being investigated. Our second task during the proposed study is to link the contextual database to the individual-level data collected under the core application and assess the separate and combined effects of school context and geographic context on each of the health behaviors. We will examine not only the "total" effects of the characteristics of both school context and geographic context, but also determine at which level of aggregation specific characteristics have the greatest impact on the health behaviors considered. Further, we will investigate whether individuals with specific attributes are more or less vulnerable to the effects of context, and examine the individual-level mechanisms (knowledge and attitudes) through which context affects decisions about engaging in health behaviors. Thus, in successfully completing the proposed study, we will provide the first comprehensive analysis of the effects of geographic and school context on a wide range of adolescent health behaviors.